Background: We established the association between arsenic ingestion and increased risk of vascular diseases associated with diabetes. However, the specific microvascular diseases in relation to arsenic exposure level have not been well reported.

Methods: The study population was obtained through national health database linkage. A total of 28 499 subjects living in the study area were successfully linked with their medical records from the National Health Insurance database in 1999-2000. The arsenic concentrations of artesian well water in the villages of the study area were utilized as indices of previous ingestion level. Both stratified analysis and unconditional logistic regression were used to examine mainly neurological and renal disease in relation to the arsenic exposure taking into account diabetes status.

Results: The age-adjusted and gender-adjusted prevalence of microvascular diseases was 7.51% [95% confidence interval (CI) 7.50-7.51] for arsenic level of <0.1 mg/litre, and then increased from 6.59% (6.59-6.60) for the arsenic concentrations of 0.1-0.29 mg/litre to 8.02% (8.02-8.03) and 11.82% (11.81-11.83) for those of 0.3-0.59 mg/litre and > or =0.6 mg/litre in non-diabetic subjects. For diabetic patients, the prevalence was 16.41% (95% CI 16.37-16.45), 15.85% (15.8-15.9), 21.69% (21.6-21.8), and 28.31% (28.2-28.4) for arsenic levels <0.1, 0.1-0.29, 0.3-0.59, and > or =0.6 mg/litre, respectively. The prevalence of microvascular diseases increased significantly with arsenic exposure, especially at higher levels, and the relationship is stronger in diabetics than in non-diabetic subjects. The results for neurological disease are very similar, and the patterns are the same for renal disease.

Conclusions: The increased prevalence of microvascular diseases, including neurological and renal disorders, is associated with arsenic ingestion. This excess risk was further elevated in diabetic subjects. Further studies are necessary to verify the hypotheses of threshold or dose-response relationships.